Mimir query engine: add support for aggregations with `without`, and resolve potential for hash collisions #8671

charleskorn · 2024-07-10T02:53:41Z

What this PR does

This PR adds support for the without keyword for aggregations in Mimir's query engine.

It also resolves the potential for hash collisions when computing aggregation groups. This decreases performance somewhat, but is the cheapest way I can see to avoid hash collisions.

If we're happy with the seriesToGroupLabelsStringFunc approach here, I will investigate reusing it for binary operations in a separate PR.

Performance comparison of this PR vs Prometheus' engine

goos: darwin
goarch: arm64
pkg: github.com/grafana/mimir/pkg/streamingpromql/benchmarks
                                                                                                         │      Prometheus      │               Mimir                │
                                                                                                         │        sec/op        │   sec/op     vs base               │
Query/sum(a_1),_instant_query                                                                                      139.6µ ±  2%   140.3µ ± 8%        ~ (p=0.937 n=6)
Query/sum(a_1),_range_query_with_100_steps                                                                         154.2µ ±  4%   153.1µ ± 2%        ~ (p=0.937 n=6)
Query/sum(a_1),_range_query_with_1000_steps                                                                        257.0µ ±  2%   236.0µ ± 8%   -8.18% (p=0.004 n=6)
Query/sum(a_100),_instant_query                                                                                    851.1µ ±  2%   855.9µ ± 1%        ~ (p=0.485 n=6)
Query/sum(a_100),_range_query_with_100_steps                                                                       1.497m ±  1%   1.438m ± 3%   -3.95% (p=0.009 n=6)
Query/sum(a_100),_range_query_with_1000_steps                                                                      7.104m ±  0%   6.749m ± 1%   -5.00% (p=0.002 n=6)
Query/sum(a_2000),_instant_query                                                                                   11.68m ±  1%   11.39m ± 2%   -2.52% (p=0.009 n=6)
Query/sum(a_2000),_range_query_with_100_steps                                                                      22.93m ±  1%   21.83m ± 1%   -4.76% (p=0.002 n=6)
Query/sum(a_2000),_range_query_with_1000_steps                                                                     139.1m ±  1%   118.3m ± 1%  -14.99% (p=0.002 n=6)
Query/sum_without_(l)(h_1),_instant_query                                                                          171.9µ ±  1%   169.8µ ± 2%   -1.20% (p=0.002 n=6)
Query/sum_without_(l)(h_1),_range_query_with_100_steps                                                             220.1µ ±  1%   213.4µ ± 2%   -3.03% (p=0.002 n=6)
Query/sum_without_(l)(h_1),_range_query_with_1000_steps                                                            666.1µ ±  7%   599.9µ ± 2%   -9.94% (p=0.002 n=6)
Query/sum_without_(l)(h_100),_instant_query                                                                        4.035m ±  3%   3.943m ± 0%   -2.28% (p=0.002 n=6)
Query/sum_without_(l)(h_100),_range_query_with_100_steps                                                           7.641m ±  1%   7.347m ± 1%   -3.86% (p=0.002 n=6)
Query/sum_without_(l)(h_100),_range_query_with_1000_steps                                                          41.28m ±  3%   38.40m ± 0%   -6.97% (p=0.002 n=6)
Query/sum_without_(l)(h_2000),_instant_query                                                                       74.57m ±  1%   72.93m ± 0%   -2.20% (p=0.002 n=6)
Query/sum_without_(l)(h_2000),_range_query_with_100_steps                                                          147.0m ±  1%   139.9m ± 1%   -4.85% (p=0.002 n=6)
Query/sum_without_(l)(h_2000),_range_query_with_1000_steps                                                         875.9m ±  0%   735.7m ± 1%  -16.00% (p=0.002 n=6)
Query/sum_without_(le)(h_1),_instant_query                                                                         170.1µ ±  1%   166.8µ ± 6%        ~ (p=0.394 n=6)
Query/sum_without_(le)(h_1),_range_query_with_100_steps                                                            215.6µ ±  0%   210.9µ ± 4%        ~ (p=0.065 n=6)
Query/sum_without_(le)(h_1),_range_query_with_1000_steps                                                           632.5µ ±  1%   588.7µ ± 6%   -6.93% (p=0.002 n=6)
Query/sum_without_(le)(h_100),_instant_query                                                                       4.082m ±  1%   3.952m ± 0%   -3.19% (p=0.002 n=6)
Query/sum_without_(le)(h_100),_range_query_with_100_steps                                                          7.873m ±  2%   7.415m ± 4%   -5.82% (p=0.002 n=6)
Query/sum_without_(le)(h_100),_range_query_with_1000_steps                                                         42.48m ±  1%   38.33m ± 1%   -9.77% (p=0.002 n=6)
Query/sum_without_(le)(h_2000),_instant_query                                                                      75.44m ±  2%   74.67m ± 1%        ~ (p=0.132 n=6)
Query/sum_without_(le)(h_2000),_range_query_with_100_steps                                                         150.5m ±  1%   140.0m ± 1%   -6.94% (p=0.002 n=6)
Query/sum_without_(le)(h_2000),_range_query_with_1000_steps                                                        897.8m ±  0%   735.8m ± 1%  -18.05% (p=0.002 n=6)
Query/sum_by_(l)(h_1),_instant_query                                                                               169.9µ ±  5%   167.8µ ± 6%        ~ (p=0.310 n=6)
Query/sum_by_(l)(h_1),_range_query_with_100_steps                                                                  215.5µ ±  1%   209.7µ ± 4%        ~ (p=0.065 n=6)
Query/sum_by_(l)(h_1),_range_query_with_1000_steps                                                                 633.4µ ±  0%   586.6µ ± 1%   -7.39% (p=0.002 n=6)
Query/sum_by_(l)(h_100),_instant_query                                                                             4.101m ±  0%   3.949m ± 0%   -3.69% (p=0.002 n=6)
Query/sum_by_(l)(h_100),_range_query_with_100_steps                                                                7.790m ±  1%   7.383m ± 4%   -5.23% (p=0.002 n=6)
Query/sum_by_(l)(h_100),_range_query_with_1000_steps                                                               42.30m ±  2%   38.65m ± 1%   -8.62% (p=0.002 n=6)
Query/sum_by_(l)(h_2000),_instant_query                                                                            74.57m ±  1%   74.18m ± 1%        ~ (p=0.240 n=6)
Query/sum_by_(l)(h_2000),_range_query_with_100_steps                                                               151.6m ±  2%   140.3m ± 1%   -7.40% (p=0.002 n=6)
Query/sum_by_(l)(h_2000),_range_query_with_1000_steps                                                              904.2m ±  3%   735.0m ± 1%  -18.71% (p=0.002 n=6)
Query/sum_by_(le)(h_1),_instant_query                                                                              172.0µ ±  1%   168.5µ ± 1%   -2.02% (p=0.002 n=6)
Query/sum_by_(le)(h_1),_range_query_with_100_steps                                                                 236.0µ ±  6%   212.8µ ± 2%   -9.81% (p=0.002 n=6)
Query/sum_by_(le)(h_1),_range_query_with_1000_steps                                                                694.3µ ±  7%   603.0µ ± 4%  -13.16% (p=0.002 n=6)
Query/sum_by_(le)(h_100),_instant_query                                                                            4.180m ±  1%   3.955m ± 1%   -5.38% (p=0.002 n=6)
Query/sum_by_(le)(h_100),_range_query_with_100_steps                                                               7.851m ±  1%   7.448m ± 2%   -5.13% (p=0.002 n=6)
Query/sum_by_(le)(h_100),_range_query_with_1000_steps                                                              42.18m ±  5%   38.68m ± 1%   -8.29% (p=0.002 n=6)
Query/sum_by_(le)(h_2000),_instant_query                                                                           75.80m ±  5%   73.05m ± 0%   -3.62% (p=0.002 n=6)
Query/sum_by_(le)(h_2000),_range_query_with_100_steps                                                              149.0m ±  3%   140.2m ± 0%   -5.88% (p=0.002 n=6)
Query/sum_by_(le)(h_2000),_range_query_with_1000_steps                                                             894.4m ±  2%   740.7m ± 0%  -17.19% (p=0.002 n=6)
Query/sum(a_1_+_b_1),_instant_query                                                                                249.9µ ±  6%   230.6µ ± 4%   -7.73% (p=0.002 n=6)
Query/sum(a_1_+_b_1),_range_query_with_100_steps                                                                   293.3µ ±  4%   250.7µ ± 1%  -14.53% (p=0.002 n=6)
Query/sum(a_1_+_b_1),_range_query_with_1000_steps                                                                  757.4µ ± 12%   395.3µ ± 3%  -47.81% (p=0.002 n=6)
Query/sum(a_100_+_b_100),_instant_query                                                                            1.640m ±  1%   1.605m ± 1%   -2.12% (p=0.002 n=6)
Query/sum(a_100_+_b_100),_range_query_with_100_steps                                                               5.164m ±  1%   2.745m ± 1%  -46.85% (p=0.002 n=6)
Query/sum(a_100_+_b_100),_range_query_with_1000_steps                                                              37.37m ±  2%   13.11m ± 1%  -64.92% (p=0.002 n=6)
Query/sum(a_2000_+_b_2000),_instant_query                                                                          25.01m ±  1%   25.00m ± 1%        ~ (p=0.818 n=6)
Query/sum(a_2000_+_b_2000),_range_query_with_100_steps                                                             96.59m ±  2%   46.06m ± 1%  -52.31% (p=0.002 n=6)
Query/sum(a_2000_+_b_2000),_range_query_with_1000_steps                                                            804.1m ±  0%   243.1m ± 0%  -69.77% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_1[1m])),_instant_query                                                                    149.5µ ±  5%   135.9µ ± 1%   -9.09% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_1[1m])),_range_query_with_100_steps                                                       245.6µ ±  2%   187.4µ ± 1%  -23.68% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_1[1m])),_range_query_with_1000_steps                                                     1009.2µ ±  3%   634.9µ ± 8%  -37.09% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_100[1m])),_instant_query                                                                  2.085m ±  1%   1.583m ± 3%  -24.09% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_100[1m])),_range_query_with_100_steps                                                     8.970m ±  8%   5.644m ± 1%  -37.08% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_100[1m])),_range_query_with_1000_steps                                                    76.44m ±  2%   42.31m ± 1%  -44.65% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_2000[1m])),_instant_query                                                                 36.94m ±  1%   26.43m ± 1%  -28.45% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_2000[1m])),_range_query_with_100_steps                                                    182.5m ±  4%   105.9m ± 1%  -41.98% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_2000[1m])),_range_query_with_1000_steps                                                  1749.1m ± 13%   808.7m ± 1%  -53.77% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_1[1m])),_instant_query                                                                123.5µ ± 14%   117.9µ ± 5%        ~ (p=0.065 n=6)
Query/sum_without_(l)(rate(a_1[1m])),_range_query_with_100_steps                                                   147.3µ ±  3%   131.1µ ± 7%  -10.99% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_1[1m])),_range_query_with_1000_steps                                                  306.7µ ±  5%   220.8µ ± 6%  -28.01% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_100[1m])),_instant_query                                                              514.1µ ±  1%   374.1µ ± 7%  -27.23% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_100[1m])),_range_query_with_100_steps                                                 1.720m ±  5%   1.115m ± 1%  -35.14% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_100[1m])),_range_query_with_1000_steps                                               13.152m ±  7%   7.303m ± 1%  -44.47% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_2000[1m])),_instant_query                                                             6.025m ±  1%   4.112m ± 1%  -31.75% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_2000[1m])),_range_query_with_100_steps                                                30.69m ±  7%   16.96m ± 0%  -44.73% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_2000[1m])),_range_query_with_1000_steps                                               260.5m ±  7%   131.2m ± 1%  -49.62% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_1[1m]))_/_sum_without_(l)(rate(b_1[1m])),_instant_query                               237.6µ ±  4%   223.4µ ± 1%   -5.96% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_1[1m]))_/_sum_without_(l)(rate(b_1[1m])),_range_query_with_100_steps                  294.7µ ±  1%   247.6µ ± 1%  -15.99% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_1[1m]))_/_sum_without_(l)(rate(b_1[1m])),_range_query_with_1000_steps                 859.1µ ±  1%   417.1µ ± 1%  -51.45% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_100[1m]))_/_sum_without_(l)(rate(b_100[1m])),_instant_query                           993.3µ ±  4%   779.6µ ± 6%  -21.52% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_100[1m]))_/_sum_without_(l)(rate(b_100[1m])),_range_query_with_100_steps              3.291m ±  3%   2.113m ± 1%  -35.81% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_100[1m]))_/_sum_without_(l)(rate(b_100[1m])),_range_query_with_1000_steps             23.71m ±  3%   13.71m ± 1%  -42.19% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_2000[1m]))_/_sum_without_(l)(rate(b_2000[1m])),_instant_query                        11.872m ±  1%   8.231m ± 0%  -30.67% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_2000[1m]))_/_sum_without_(l)(rate(b_2000[1m])),_range_query_with_100_steps            57.45m ±  5%   34.17m ± 1%  -40.53% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_2000[1m]))_/_sum_without_(l)(rate(b_2000[1m])),_range_query_with_1000_steps           482.3m ±  0%   261.5m ± 1%  -45.78% (p=0.002 n=6)
geomean                                                                                                            6.060m         4.830m       -20.30%


                                                                                                         │      Prometheus      │                Mimir                │
                                                                                                         │          B           │      B        vs base               │
Query/sum(a_1),_instant_query                                                                                      71.81Mi ± 1%   71.73Mi ± 2%        ~ (p=0.974 n=6)
Query/sum(a_1),_range_query_with_100_steps                                                                         71.50Mi ± 2%   70.77Mi ± 1%        ~ (p=0.485 n=6)
Query/sum(a_1),_range_query_with_1000_steps                                                                        70.64Mi ± 2%   69.50Mi ± 1%   -1.61% (p=0.041 n=6)
Query/sum(a_100),_instant_query                                                                                    65.43Mi ± 1%   64.55Mi ± 1%   -1.35% (p=0.002 n=6)
Query/sum(a_100),_range_query_with_100_steps                                                                       65.21Mi ± 1%   64.70Mi ± 1%   -0.79% (p=0.002 n=6)
Query/sum(a_100),_range_query_with_1000_steps                                                                      67.97Mi ± 1%   64.56Mi ± 1%   -5.01% (p=0.002 n=6)
Query/sum(a_2000),_instant_query                                                                                   68.18Mi ± 4%   66.12Mi ± 2%   -3.03% (p=0.041 n=6)
Query/sum(a_2000),_range_query_with_100_steps                                                                      73.54Mi ± 3%   64.83Mi ± 1%  -11.85% (p=0.002 n=6)
Query/sum(a_2000),_range_query_with_1000_steps                                                                    130.39Mi ± 1%   69.26Mi ± 4%  -46.88% (p=0.002 n=6)
Query/sum_without_(l)(h_1),_instant_query                                                                          69.23Mi ± 1%   68.69Mi ± 1%        ~ (p=0.065 n=6)
Query/sum_without_(l)(h_1),_range_query_with_100_steps                                                             69.38Mi ± 2%   68.23Mi ± 0%        ~ (p=0.132 n=6)
Query/sum_without_(l)(h_1),_range_query_with_1000_steps                                                            66.82Mi ± 1%   66.55Mi ± 1%        ~ (p=0.255 n=6)
Query/sum_without_(l)(h_100),_instant_query                                                                        66.13Mi ± 1%   65.65Mi ± 1%   -0.73% (p=0.006 n=6)
Query/sum_without_(l)(h_100),_range_query_with_100_steps                                                           68.21Mi ± 1%   65.76Mi ± 1%   -3.60% (p=0.002 n=6)
Query/sum_without_(l)(h_100),_range_query_with_1000_steps                                                          85.62Mi ± 1%   66.23Mi ± 2%  -22.65% (p=0.002 n=6)
Query/sum_without_(l)(h_2000),_instant_query                                                                       70.76Mi ± 2%   69.82Mi ± 0%        ~ (p=0.394 n=6)
Query/sum_without_(l)(h_2000),_range_query_with_100_steps                                                         110.44Mi ± 1%   69.65Mi ± 1%  -36.93% (p=0.002 n=6)
Query/sum_without_(l)(h_2000),_range_query_with_1000_steps                                                        406.73Mi ± 0%   72.62Mi ± 1%  -82.14% (p=0.002 n=6)
Query/sum_without_(le)(h_1),_instant_query                                                                         69.62Mi ± 1%   69.16Mi ± 1%   -0.66% (p=0.002 n=6)
Query/sum_without_(le)(h_1),_range_query_with_100_steps                                                            68.98Mi ± 2%   68.23Mi ± 1%   -1.09% (p=0.002 n=6)
Query/sum_without_(le)(h_1),_range_query_with_1000_steps                                                           67.05Mi ± 1%   66.14Mi ± 2%   -1.36% (p=0.026 n=6)
Query/sum_without_(le)(h_100),_instant_query                                                                       66.03Mi ± 1%   65.73Mi ± 1%        ~ (p=0.058 n=6)
Query/sum_without_(le)(h_100),_range_query_with_100_steps                                                          68.55Mi ± 1%   66.38Mi ± 1%   -3.16% (p=0.002 n=6)
Query/sum_without_(le)(h_100),_range_query_with_1000_steps                                                         87.92Mi ± 2%   68.95Mi ± 1%  -21.57% (p=0.002 n=6)
Query/sum_without_(le)(h_2000),_instant_query                                                                      71.56Mi ± 2%   70.43Mi ± 2%        ~ (p=0.108 n=6)
Query/sum_without_(le)(h_2000),_range_query_with_100_steps                                                        118.12Mi ± 1%   77.95Mi ± 1%  -34.01% (p=0.002 n=6)
Query/sum_without_(le)(h_2000),_range_query_with_1000_steps                                                        439.8Mi ± 0%   131.2Mi ± 5%  -70.18% (p=0.002 n=6)
Query/sum_by_(l)(h_1),_instant_query                                                                               69.77Mi ± 1%   68.64Mi ± 1%   -1.62% (p=0.009 n=6)
Query/sum_by_(l)(h_1),_range_query_with_100_steps                                                                  69.02Mi ± 1%   68.22Mi ± 1%   -1.15% (p=0.004 n=6)
Query/sum_by_(l)(h_1),_range_query_with_1000_steps                                                                 66.88Mi ± 1%   66.11Mi ± 2%   -1.14% (p=0.039 n=6)
Query/sum_by_(l)(h_100),_instant_query                                                                             66.19Mi ± 1%   65.72Mi ± 1%   -0.71% (p=0.041 n=6)
Query/sum_by_(l)(h_100),_range_query_with_100_steps                                                                68.67Mi ± 2%   66.15Mi ± 2%   -3.67% (p=0.002 n=6)
Query/sum_by_(l)(h_100),_range_query_with_1000_steps                                                               87.01Mi ± 1%   69.30Mi ± 2%  -20.35% (p=0.002 n=6)
Query/sum_by_(l)(h_2000),_instant_query                                                                            70.12Mi ± 3%   70.46Mi ± 2%        ~ (p=1.000 n=6)
Query/sum_by_(l)(h_2000),_range_query_with_100_steps                                                              117.10Mi ± 1%   78.97Mi ± 3%  -32.56% (p=0.002 n=6)
Query/sum_by_(l)(h_2000),_range_query_with_1000_steps                                                              439.8Mi ± 0%   133.7Mi ± 4%  -69.61% (p=0.002 n=6)
Query/sum_by_(le)(h_1),_instant_query                                                                              69.60Mi ± 1%   68.97Mi ± 1%        ~ (p=0.061 n=6)
Query/sum_by_(le)(h_1),_range_query_with_100_steps                                                                 68.93Mi ± 1%   68.26Mi ± 1%   -0.97% (p=0.002 n=6)
Query/sum_by_(le)(h_1),_range_query_with_1000_steps                                                                67.19Mi ± 1%   65.93Mi ± 1%   -1.87% (p=0.002 n=6)
Query/sum_by_(le)(h_100),_instant_query                                                                            66.27Mi ± 2%   65.77Mi ± 1%        ~ (p=0.180 n=6)
Query/sum_by_(le)(h_100),_range_query_with_100_steps                                                               68.27Mi ± 2%   65.59Mi ± 2%   -3.91% (p=0.002 n=6)
Query/sum_by_(le)(h_100),_range_query_with_1000_steps                                                              85.45Mi ± 4%   67.19Mi ± 2%  -21.37% (p=0.002 n=6)
Query/sum_by_(le)(h_2000),_instant_query                                                                           70.65Mi ± 2%   69.36Mi ± 1%   -1.82% (p=0.026 n=6)
Query/sum_by_(le)(h_2000),_range_query_with_100_steps                                                             110.01Mi ± 1%   69.55Mi ± 1%  -36.78% (p=0.002 n=6)
Query/sum_by_(le)(h_2000),_range_query_with_1000_steps                                                            406.40Mi ± 0%   71.93Mi ± 2%  -82.30% (p=0.002 n=6)
Query/sum(a_1_+_b_1),_instant_query                                                                                71.78Mi ± 1%   71.98Mi ± 1%        ~ (p=0.240 n=6)
Query/sum(a_1_+_b_1),_range_query_with_100_steps                                                                   71.25Mi ± 0%   71.16Mi ± 1%        ~ (p=0.558 n=6)
Query/sum(a_1_+_b_1),_range_query_with_1000_steps                                                                  67.44Mi ± 2%   69.51Mi ± 2%   +3.07% (p=0.002 n=6)
Query/sum(a_100_+_b_100),_instant_query                                                                            65.53Mi ± 0%   65.34Mi ± 1%        ~ (p=0.407 n=6)
Query/sum(a_100_+_b_100),_range_query_with_100_steps                                                               66.25Mi ± 1%   65.04Mi ± 1%   -1.83% (p=0.002 n=6)
Query/sum(a_100_+_b_100),_range_query_with_1000_steps                                                              73.03Mi ± 2%   64.46Mi ± 1%  -11.74% (p=0.002 n=6)
Query/sum(a_2000_+_b_2000),_instant_query                                                                          68.58Mi ± 1%   67.93Mi ± 2%        ~ (p=0.394 n=6)
Query/sum(a_2000_+_b_2000),_range_query_with_100_steps                                                             89.65Mi ± 1%   68.63Mi ± 3%  -23.44% (p=0.002 n=6)
Query/sum(a_2000_+_b_2000),_range_query_with_1000_steps                                                           204.57Mi ± 0%   79.00Mi ± 4%  -61.38% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_1[1m])),_instant_query                                                                    69.29Mi ± 1%   69.24Mi ± 1%        ~ (p=0.665 n=6)
Query/sum_by_(le)(rate(h_1[1m])),_range_query_with_100_steps                                                       68.84Mi ± 1%   68.84Mi ± 1%        ~ (p=0.457 n=6)
Query/sum_by_(le)(rate(h_1[1m])),_range_query_with_1000_steps                                                      66.46Mi ± 1%   65.89Mi ± 1%        ~ (p=0.082 n=6)
Query/sum_by_(le)(rate(h_100[1m])),_instant_query                                                                  66.41Mi ± 1%   66.02Mi ± 2%        ~ (p=0.119 n=6)
Query/sum_by_(le)(rate(h_100[1m])),_range_query_with_100_steps                                                     67.52Mi ± 1%   65.17Mi ± 2%   -3.48% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_100[1m])),_range_query_with_1000_steps                                                    83.19Mi ± 3%   66.12Mi ± 2%  -20.52% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_2000[1m])),_instant_query                                                                 72.14Mi ± 1%   71.88Mi ± 1%        ~ (p=0.589 n=6)
Query/sum_by_(le)(rate(h_2000[1m])),_range_query_with_100_steps                                                   108.57Mi ± 3%   69.08Mi ± 1%  -36.37% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_2000[1m])),_range_query_with_1000_steps                                                  268.34Mi ± 1%   71.63Mi ± 2%  -73.30% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_1[1m])),_instant_query                                                                71.50Mi ± 0%   71.86Mi ± 2%        ~ (p=0.180 n=6)
Query/sum_without_(l)(rate(a_1[1m])),_range_query_with_100_steps                                                   70.94Mi ± 2%   71.15Mi ± 1%        ~ (p=0.394 n=6)
Query/sum_without_(l)(rate(a_1[1m])),_range_query_with_1000_steps                                                  69.30Mi ± 1%   69.31Mi ± 2%        ~ (p=0.784 n=6)
Query/sum_without_(l)(rate(a_100[1m])),_instant_query                                                              65.94Mi ± 1%   65.20Mi ± 1%        ~ (p=0.240 n=6)
Query/sum_without_(l)(rate(a_100[1m])),_range_query_with_100_steps                                                 65.40Mi ± 2%   64.43Mi ± 1%   -1.48% (p=0.024 n=6)
Query/sum_without_(l)(rate(a_100[1m])),_range_query_with_1000_steps                                                65.48Mi ± 2%   64.20Mi ± 1%   -1.97% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_2000[1m])),_instant_query                                                             68.44Mi ± 3%   66.73Mi ± 1%        ~ (p=0.132 n=6)
Query/sum_without_(l)(rate(a_2000[1m])),_range_query_with_100_steps                                                72.77Mi ± 2%   65.38Mi ± 1%  -10.16% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_2000[1m])),_range_query_with_1000_steps                                              125.64Mi ± 3%   68.52Mi ± 2%  -45.46% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_1[1m]))_/_sum_without_(l)(rate(b_1[1m])),_instant_query                               71.12Mi ± 1%   71.23Mi ± 1%        ~ (p=0.394 n=6)
Query/sum_without_(l)(rate(a_1[1m]))_/_sum_without_(l)(rate(b_1[1m])),_range_query_with_100_steps                  70.20Mi ± 1%   70.70Mi ± 1%        ~ (p=0.310 n=6)
Query/sum_without_(l)(rate(a_1[1m]))_/_sum_without_(l)(rate(b_1[1m])),_range_query_with_1000_steps                 67.70Mi ± 1%   68.72Mi ± 1%   +1.50% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_100[1m]))_/_sum_without_(l)(rate(b_100[1m])),_instant_query                           66.11Mi ± 1%   65.69Mi ± 0%   -0.64% (p=0.013 n=6)
Query/sum_without_(l)(rate(a_100[1m]))_/_sum_without_(l)(rate(b_100[1m])),_range_query_with_100_steps              65.38Mi ± 2%   64.89Mi ± 0%   -0.74% (p=0.015 n=6)
Query/sum_without_(l)(rate(a_100[1m]))_/_sum_without_(l)(rate(b_100[1m])),_range_query_with_1000_steps             66.41Mi ± 2%   64.51Mi ± 3%   -2.87% (p=0.009 n=6)
Query/sum_without_(l)(rate(a_2000[1m]))_/_sum_without_(l)(rate(b_2000[1m])),_instant_query                         67.52Mi ± 1%   67.12Mi ± 1%        ~ (p=0.394 n=6)
Query/sum_without_(l)(rate(a_2000[1m]))_/_sum_without_(l)(rate(b_2000[1m])),_range_query_with_100_steps            74.07Mi ± 1%   66.74Mi ± 2%   -9.89% (p=0.002 n=6)
Query/sum_without_(l)(rate(a_2000[1m]))_/_sum_without_(l)(rate(b_2000[1m])),_range_query_with_1000_steps          127.55Mi ± 1%   74.03Mi ± 2%  -41.96% (p=0.002 n=6)
geomean                                                                                                            83.10Mi        69.35Mi       -16.54%

Performance comparison of this PR with main

goos: darwin
goarch: arm64
pkg: github.com/grafana/mimir/pkg/streamingpromql/benchmarks
                                                                          │  main.txt   │  after-with-no-string-alloc.txt   │
                                                                          │   sec/op    │   sec/op     vs base              │
Query/sum(a_1),_instant_query/Mimir-10                                      140.2µ ± 2%   140.3µ ± 8%       ~ (p=1.000 n=6)
Query/sum(a_1),_range_query_with_100_steps/Mimir-10                         153.5µ ± 2%   153.1µ ± 2%       ~ (p=0.818 n=6)
Query/sum(a_1),_range_query_with_1000_steps/Mimir-10                        232.7µ ± 1%   236.0µ ± 8%       ~ (p=0.485 n=6)
Query/sum(a_100),_instant_query/Mimir-10                                    843.1µ ± 1%   855.9µ ± 1%  +1.51% (p=0.009 n=6)
Query/sum(a_100),_range_query_with_100_steps/Mimir-10                       1.470m ± 3%   1.438m ± 3%       ~ (p=0.180 n=6)
Query/sum(a_100),_range_query_with_1000_steps/Mimir-10                      6.714m ± 1%   6.749m ± 1%       ~ (p=0.394 n=6)
Query/sum(a_2000),_instant_query/Mimir-10                                   11.51m ± 1%   11.39m ± 2%       ~ (p=0.180 n=6)
Query/sum(a_2000),_range_query_with_100_steps/Mimir-10                      22.08m ± 1%   21.83m ± 1%       ~ (p=0.065 n=6)
Query/sum(a_2000),_range_query_with_1000_steps/Mimir-10                     119.4m ± 1%   118.3m ± 1%       ~ (p=0.065 n=6)
Query/sum_by_(l)(h_1),_instant_query/Mimir-10                               166.2µ ± 1%   167.8µ ± 6%       ~ (p=0.240 n=6)
Query/sum_by_(l)(h_1),_range_query_with_100_steps/Mimir-10                  206.0µ ± 1%   209.7µ ± 4%  +1.76% (p=0.026 n=6)
Query/sum_by_(l)(h_1),_range_query_with_1000_steps/Mimir-10                 588.0µ ± 2%   586.6µ ± 1%       ~ (p=0.818 n=6)
Query/sum_by_(l)(h_100),_instant_query/Mimir-10                             3.991m ± 1%   3.949m ± 0%  -1.05% (p=0.002 n=6)
Query/sum_by_(l)(h_100),_range_query_with_100_steps/Mimir-10                7.433m ± 1%   7.383m ± 4%       ~ (p=0.589 n=6)
Query/sum_by_(l)(h_100),_range_query_with_1000_steps/Mimir-10               38.35m ± 1%   38.65m ± 1%       ~ (p=0.065 n=6)
Query/sum_by_(l)(h_2000),_instant_query/Mimir-10                            74.60m ± 1%   74.18m ± 1%  -0.57% (p=0.041 n=6)
Query/sum_by_(l)(h_2000),_range_query_with_100_steps/Mimir-10               140.2m ± 1%   140.3m ± 1%       ~ (p=0.937 n=6)
Query/sum_by_(l)(h_2000),_range_query_with_1000_steps/Mimir-10              734.2m ± 0%   735.0m ± 1%       ~ (p=1.000 n=6)
Query/sum_by_(le)(h_1),_instant_query/Mimir-10                              165.9µ ± 7%   168.5µ ± 1%       ~ (p=0.065 n=6)
Query/sum_by_(le)(h_1),_range_query_with_100_steps/Mimir-10                 208.4µ ± 1%   212.8µ ± 2%  +2.14% (p=0.002 n=6)
Query/sum_by_(le)(h_1),_range_query_with_1000_steps/Mimir-10                591.3µ ± 2%   603.0µ ± 4%  +1.97% (p=0.041 n=6)
Query/sum_by_(le)(h_100),_instant_query/Mimir-10                            3.952m ± 0%   3.955m ± 1%       ~ (p=0.937 n=6)
Query/sum_by_(le)(h_100),_range_query_with_100_steps/Mimir-10               7.364m ± 0%   7.448m ± 2%       ~ (p=0.065 n=6)
Query/sum_by_(le)(h_100),_range_query_with_1000_steps/Mimir-10              38.58m ± 0%   38.68m ± 1%       ~ (p=0.394 n=6)
Query/sum_by_(le)(h_2000),_instant_query/Mimir-10                           73.47m ± 0%   73.05m ± 0%  -0.56% (p=0.004 n=6)
Query/sum_by_(le)(h_2000),_range_query_with_100_steps/Mimir-10              141.0m ± 1%   140.2m ± 0%       ~ (p=0.699 n=6)
Query/sum_by_(le)(h_2000),_range_query_with_1000_steps/Mimir-10             737.8m ± 1%   740.7m ± 0%       ~ (p=0.093 n=6)
Query/sum(a_1_+_b_1),_instant_query/Mimir-10                                229.9µ ± 1%   230.6µ ± 4%       ~ (p=0.394 n=6)
Query/sum(a_1_+_b_1),_range_query_with_100_steps/Mimir-10                   249.2µ ± 1%   250.7µ ± 1%  +0.58% (p=0.041 n=6)
Query/sum(a_1_+_b_1),_range_query_with_1000_steps/Mimir-10                  388.1µ ± 1%   395.3µ ± 3%  +1.85% (p=0.002 n=6)
Query/sum(a_100_+_b_100),_instant_query/Mimir-10                            1.607m ± 1%   1.605m ± 1%       ~ (p=0.310 n=6)
Query/sum(a_100_+_b_100),_range_query_with_100_steps/Mimir-10               2.744m ± 1%   2.745m ± 1%       ~ (p=0.818 n=6)
Query/sum(a_100_+_b_100),_range_query_with_1000_steps/Mimir-10              13.11m ± 0%   13.11m ± 1%       ~ (p=0.937 n=6)
Query/sum(a_2000_+_b_2000),_instant_query/Mimir-10                          24.99m ± 1%   25.00m ± 1%       ~ (p=0.937 n=6)
Query/sum(a_2000_+_b_2000),_range_query_with_100_steps/Mimir-10             46.32m ± 1%   46.06m ± 1%       ~ (p=0.394 n=6)
Query/sum(a_2000_+_b_2000),_range_query_with_1000_steps/Mimir-10            243.8m ± 0%   243.1m ± 0%       ~ (p=0.065 n=6)
Query/sum_by_(le)(rate(h_1[1m])),_instant_query/Mimir-10                    134.4µ ± 1%   135.9µ ± 1%  +1.09% (p=0.041 n=6)
Query/sum_by_(le)(rate(h_1[1m])),_range_query_with_100_steps/Mimir-10       187.6µ ± 1%   187.4µ ± 1%       ~ (p=0.937 n=6)
Query/sum_by_(le)(rate(h_1[1m])),_range_query_with_1000_steps/Mimir-10      629.9µ ± 1%   634.9µ ± 8%  +0.80% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_100[1m])),_instant_query/Mimir-10                  1.513m ± 1%   1.583m ± 3%  +4.63% (p=0.002 n=6)
Query/sum_by_(le)(rate(h_100[1m])),_range_query_with_100_steps/Mimir-10     5.606m ± 1%   5.644m ± 1%       ~ (p=0.065 n=6)
Query/sum_by_(le)(rate(h_100[1m])),_range_query_with_1000_steps/Mimir-10    42.41m ± 0%   42.31m ± 1%       ~ (p=0.180 n=6)
Query/sum_by_(le)(rate(h_2000[1m])),_instant_query/Mimir-10                 26.04m ± 1%   26.43m ± 1%  +1.51% (p=0.004 n=6)
Query/sum_by_(le)(rate(h_2000[1m])),_range_query_with_100_steps/Mimir-10    106.7m ± 1%   105.9m ± 1%  -0.80% (p=0.041 n=6)
Query/sum_by_(le)(rate(h_2000[1m])),_range_query_with_1000_steps/Mimir-10   816.8m ± 1%   808.7m ± 1%  -0.99% (p=0.004 n=6)
geomean                                                                     5.212m        5.227m       +0.30%

                                                                          │   main.txt   │   after-with-no-string-alloc.txt   │
                                                                          │      B       │      B        vs base              │
Query/sum(a_1),_instant_query/Mimir-10                                      71.38Mi ± 1%   71.73Mi ± 2%       ~ (p=0.619 n=6)
Query/sum(a_1),_range_query_with_100_steps/Mimir-10                         70.22Mi ± 1%   70.77Mi ± 1%  +0.78% (p=0.026 n=6)
Query/sum(a_1),_range_query_with_1000_steps/Mimir-10                        68.88Mi ± 1%   69.50Mi ± 1%       ~ (p=0.065 n=6)
Query/sum(a_100),_instant_query/Mimir-10                                    64.41Mi ± 0%   64.55Mi ± 1%       ~ (p=0.509 n=6)
Query/sum(a_100),_range_query_with_100_steps/Mimir-10                       64.23Mi ± 1%   64.70Mi ± 1%  +0.72% (p=0.035 n=6)
Query/sum(a_100),_range_query_with_1000_steps/Mimir-10                      64.58Mi ± 2%   64.56Mi ± 1%       ~ (p=0.589 n=6)
Query/sum(a_2000),_instant_query/Mimir-10                                   65.60Mi ± 1%   66.12Mi ± 2%       ~ (p=0.485 n=6)
Query/sum(a_2000),_range_query_with_100_steps/Mimir-10                      65.18Mi ± 2%   64.83Mi ± 1%       ~ (p=0.240 n=6)
Query/sum(a_2000),_range_query_with_1000_steps/Mimir-10                     69.77Mi ± 1%   69.26Mi ± 4%  -0.73% (p=0.015 n=6)
Query/sum_by_(l)(h_1),_instant_query/Mimir-10                               69.31Mi ± 2%   68.64Mi ± 1%       ~ (p=0.061 n=6)
Query/sum_by_(l)(h_1),_range_query_with_100_steps/Mimir-10                  68.09Mi ± 2%   68.22Mi ± 1%       ~ (p=0.589 n=6)
Query/sum_by_(l)(h_1),_range_query_with_1000_steps/Mimir-10                 66.02Mi ± 1%   66.11Mi ± 2%       ~ (p=0.677 n=6)
Query/sum_by_(l)(h_100),_instant_query/Mimir-10                             65.34Mi ± 1%   65.72Mi ± 1%       ~ (p=0.331 n=6)
Query/sum_by_(l)(h_100),_range_query_with_100_steps/Mimir-10                66.63Mi ± 2%   66.15Mi ± 2%       ~ (p=0.310 n=6)
Query/sum_by_(l)(h_100),_range_query_with_1000_steps/Mimir-10               69.43Mi ± 1%   69.30Mi ± 2%       ~ (p=0.818 n=6)
Query/sum_by_(l)(h_2000),_instant_query/Mimir-10                            70.53Mi ± 2%   70.46Mi ± 2%       ~ (p=0.485 n=6)
Query/sum_by_(l)(h_2000),_range_query_with_100_steps/Mimir-10               78.86Mi ± 3%   78.97Mi ± 3%       ~ (p=0.818 n=6)
Query/sum_by_(l)(h_2000),_range_query_with_1000_steps/Mimir-10              135.5Mi ± 5%   133.7Mi ± 4%       ~ (p=1.000 n=6)
Query/sum_by_(le)(h_1),_instant_query/Mimir-10                              69.12Mi ± 1%   68.97Mi ± 1%       ~ (p=0.485 n=6)
Query/sum_by_(le)(h_1),_range_query_with_100_steps/Mimir-10                 68.09Mi ± 1%   68.26Mi ± 1%       ~ (p=0.818 n=6)
Query/sum_by_(le)(h_1),_range_query_with_1000_steps/Mimir-10                66.27Mi ± 1%   65.93Mi ± 1%       ~ (p=0.197 n=6)
Query/sum_by_(le)(h_100),_instant_query/Mimir-10                            65.97Mi ± 1%   65.77Mi ± 1%       ~ (p=0.368 n=6)
Query/sum_by_(le)(h_100),_range_query_with_100_steps/Mimir-10               65.66Mi ± 1%   65.59Mi ± 2%       ~ (p=0.619 n=6)
Query/sum_by_(le)(h_100),_range_query_with_1000_steps/Mimir-10              66.79Mi ± 3%   67.19Mi ± 2%       ~ (p=0.394 n=6)
Query/sum_by_(le)(h_2000),_instant_query/Mimir-10                           68.97Mi ± 2%   69.36Mi ± 1%       ~ (p=0.851 n=6)
Query/sum_by_(le)(h_2000),_range_query_with_100_steps/Mimir-10              69.45Mi ± 1%   69.55Mi ± 1%       ~ (p=0.818 n=6)
Query/sum_by_(le)(h_2000),_range_query_with_1000_steps/Mimir-10             71.46Mi ± 1%   71.93Mi ± 2%       ~ (p=0.139 n=6)
Query/sum(a_1_+_b_1),_instant_query/Mimir-10                                71.95Mi ± 1%   71.98Mi ± 1%       ~ (p=0.727 n=6)
Query/sum(a_1_+_b_1),_range_query_with_100_steps/Mimir-10                   71.51Mi ± 1%   71.16Mi ± 1%       ~ (p=0.065 n=6)
Query/sum(a_1_+_b_1),_range_query_with_1000_steps/Mimir-10                  69.72Mi ± 1%   69.51Mi ± 2%       ~ (p=0.937 n=6)
Query/sum(a_100_+_b_100),_instant_query/Mimir-10                            65.22Mi ± 1%   65.34Mi ± 1%       ~ (p=0.818 n=6)
Query/sum(a_100_+_b_100),_range_query_with_100_steps/Mimir-10               65.38Mi ± 1%   65.04Mi ± 1%       ~ (p=0.584 n=6)
Query/sum(a_100_+_b_100),_range_query_with_1000_steps/Mimir-10              64.53Mi ± 1%   64.46Mi ± 1%       ~ (p=0.937 n=6)
Query/sum(a_2000_+_b_2000),_instant_query/Mimir-10                          68.23Mi ± 2%   67.93Mi ± 2%       ~ (p=0.974 n=6)
Query/sum(a_2000_+_b_2000),_range_query_with_100_steps/Mimir-10             68.36Mi ± 1%   68.63Mi ± 3%       ~ (p=1.000 n=6)
Query/sum(a_2000_+_b_2000),_range_query_with_1000_steps/Mimir-10            77.09Mi ± 3%   79.00Mi ± 4%       ~ (p=0.240 n=6)
Query/sum_by_(le)(rate(h_1[1m])),_instant_query/Mimir-10                    69.59Mi ± 2%   69.24Mi ± 1%       ~ (p=0.699 n=6)
Query/sum_by_(le)(rate(h_1[1m])),_range_query_with_100_steps/Mimir-10       68.97Mi ± 1%   68.84Mi ± 1%       ~ (p=0.937 n=6)
Query/sum_by_(le)(rate(h_1[1m])),_range_query_with_1000_steps/Mimir-10      65.96Mi ± 0%   65.89Mi ± 1%       ~ (p=0.372 n=6)
Query/sum_by_(le)(rate(h_100[1m])),_instant_query/Mimir-10                  65.58Mi ± 1%   66.02Mi ± 2%       ~ (p=0.366 n=6)
Query/sum_by_(le)(rate(h_100[1m])),_range_query_with_100_steps/Mimir-10     65.13Mi ± 1%   65.17Mi ± 2%       ~ (p=0.485 n=6)
Query/sum_by_(le)(rate(h_100[1m])),_range_query_with_1000_steps/Mimir-10    66.51Mi ± 1%   66.12Mi ± 2%       ~ (p=0.485 n=6)
Query/sum_by_(le)(rate(h_2000[1m])),_instant_query/Mimir-10                 71.50Mi ± 1%   71.88Mi ± 1%       ~ (p=0.240 n=6)
Query/sum_by_(le)(rate(h_2000[1m])),_range_query_with_100_steps/Mimir-10    68.02Mi ± 3%   69.08Mi ± 1%       ~ (p=0.093 n=6)
Query/sum_by_(le)(rate(h_2000[1m])),_range_query_with_1000_steps/Mimir-10   71.68Mi ± 2%   71.63Mi ± 2%       ~ (p=0.937 n=6)
geomean                                                                     69.28Mi        69.34Mi       +0.09%

Which issue(s) this PR fixes or relates to

(none)

Checklist

Tests updated.
[n/a] Documentation added.
CHANGELOG.md updated - the order of entries should be [CHANGE], [FEATURE], [ENHANCEMENT], [BUGFIX].
[n/a] about-versioning.md updated with experimental features.

…ith potential hash collisions when aggregating

…aggregating to a single group

jhesketh

Nice work! Just a few nits/tweaks/questions.

pkg/streamingpromql/operators/aggregation_test.go

pkg/streamingpromql/testdata/ours/aggregators.test

pkg/streamingpromql/operators/aggregation.go

pkg/streamingpromql/operators/aggregation_test.go

pkg/streamingpromql/operators/aggregation.go

…plify logic and improve performance for case where multiple grouping labels are used

… and put related functions together

jhesketh

We should add some test cases where duplicate groups exist. For example:

sum without (env, __name__, __abc__, env) metric{}

(I used __name__ explicitly since it is added. I also used __abc__ since it'll be sorted to before __name__).

pkg/streamingpromql/operators/aggregation.go

pkg/streamingpromql/operators/aggregation_test.go

…thout` with duplicate and unsorted grouping labels

charleskorn · 2024-07-16T03:42:25Z

We should add some test cases where duplicate groups exist. For example:

sum without (env, __name__, __abc__, env) metric{}

(I used __name__ explicitly since it is added. I also used __abc__ since it'll be sorted to before __name__).

Good idea, added in fff95c4.

jhesketh

Thanks. LGTM!

There's a few tests that aren't done on both by and without (such as the __aaa__ one), but I don't think they are necessary.

Would appreciate if this can be behind #8577

# Conflicts: # CHANGELOG.md

charleskorn changed the title ~~Mimir query engine: aggregations with without~~ Mimir query engine: add support for aggregations with without, and resolve potential hash collisions Jul 10, 2024

charleskorn added 6 commits July 11, 2024 11:03

Move sorting of grouping labels into NewAggregation

9829c1b

Add support for grouping aggregations with without, and fix issue w…

abafb5c

…ith potential hash collisions when aggregating

Add changelog entry

9ed4975

Avoid calling labels.Labels.String() for every input series if we're …

b87e2b4

…aggregating to a single group

Add docstring

0efdca7

Only create labels.Labels instance when we need it.

87df177

charleskorn force-pushed the charleskorn/mqe-aggregation-with-without branch from b66fb76 to 87df177 Compare July 11, 2024 01:04

Avoid allocating strings on the heap where possible

40d3fb3

charleskorn marked this pull request as ready for review July 11, 2024 01:56

charleskorn requested a review from a team as a code owner July 11, 2024 01:56

charleskorn changed the title ~~Mimir query engine: add support for aggregations with without, and resolve potential hash collisions~~ Mimir query engine: add support for aggregations with without, and resolve potential for hash collisions Jul 11, 2024

jhesketh reviewed Jul 14, 2024

View reviewed changes

charleskorn and others added 8 commits July 15, 2024 11:06

Address PR feedback: add comment about hash collisions.

c4a2121

Address PR feedback: add comment about buffer size

d0048a2

Address PR feedback: remove unnecessary counts and use a bool instead

ce303e8

Use labels.Labels.BytesWithLabels() / BytesWithoutLabels() to sim…

a93c2d6

…plify logic and improve performance for case where multiple grouping labels are used

Address PR feedback: move labelling functions to their own functions,…

f2cc1a7

… and put related functions together

Address PR feedback: add extra test case

6885343

Address PR feedback: add additional test case

4bc8691

Merge branch 'main' into charleskorn/mqe-aggregation-with-without

d69112b

jhesketh reviewed Jul 16, 2024

View reviewed changes

pkg/streamingpromql/operators/aggregation.go Outdated Show resolved Hide resolved

pkg/streamingpromql/operators/aggregation_test.go Show resolved Hide resolved

Add note about thread safety requirements

3213a0f

charleskorn force-pushed the charleskorn/mqe-aggregation-with-without branch from a38c176 to 3213a0f Compare July 16, 2024 03:27

charleskorn added 2 commits July 16, 2024 13:28

Address PR feedback: improve method names

4a95569

Address PR feedback: add test cases for aggregations with by or `wi…

fff95c4

…thout` with duplicate and unsorted grouping labels

jhesketh approved these changes Jul 16, 2024

View reviewed changes

Merge branch 'main' into charleskorn/mqe-aggregation-with-without

8ff8c2f

# Conflicts: # CHANGELOG.md

charleskorn enabled auto-merge (squash) July 17, 2024 01:07

Merge branch 'main' into charleskorn/mqe-aggregation-with-without

d7899f5

# Conflicts: # CHANGELOG.md

charleskorn merged commit 13c11a3 into main Jul 17, 2024
29 checks passed

charleskorn deleted the charleskorn/mqe-aggregation-with-without branch July 17, 2024 01:30

charleskorn mentioned this pull request Aug 8, 2024

Mimir query engine: improve performance of binary operations #8933

Merged

1 task

charleskorn mentioned this pull request Dec 4, 2024

Mimir Query Engine #10067

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Mimir query engine: add support for aggregations with `without`, and resolve potential for hash collisions #8671

Mimir query engine: add support for aggregations with `without`, and resolve potential for hash collisions #8671

Uh oh!

charleskorn commented Jul 10, 2024 •

edited

Loading

Uh oh!

jhesketh left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jhesketh left a comment

Uh oh!

Uh oh!

Uh oh!

charleskorn commented Jul 16, 2024

Uh oh!

jhesketh left a comment

Uh oh!

Uh oh!

Uh oh!

Mimir query engine: add support for aggregations with without, and resolve potential for hash collisions #8671

Mimir query engine: add support for aggregations with without, and resolve potential for hash collisions #8671

Uh oh!

Conversation

charleskorn commented Jul 10, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What this PR does

Which issue(s) this PR fixes or relates to

Checklist

Uh oh!

jhesketh left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jhesketh left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

charleskorn commented Jul 16, 2024

Uh oh!

jhesketh left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Mimir query engine: add support for aggregations with `without`, and resolve potential for hash collisions #8671

Mimir query engine: add support for aggregations with `without`, and resolve potential for hash collisions #8671

charleskorn commented Jul 10, 2024 •

edited

Loading